Pre commit formatting (#3)

- [x] Runs formatting on staged files during commit (formatting result
will be included in the commit)

I'm still not 100% about running lint-staged long term. My main issue is
that the hook is terminated if the process running git is killed. The
result is that as a dev you're forced to whatever interface for git
you're using open during lint-staged or the commit will be arborted.
(e.g. if you commit from a pane in your terminal and close it too early
the commit won't go through)

I'm in favor of merging anyways and reverting if it becomes an issue.

.husky/pre-commit

@@ -0,0 +1,4 @@
+#!/usr/bin/env sh
+. "$(dirname "$0")/_/husky.sh"
+
+npx lint-staged

TOOLING.md

@@ -0,0 +1,11 @@
+# Tooling
+
+## Git hooks
+
+Packages:
+
+- Husky: Easier git hook configuration
+
+- Lint-staged: Run e.g. formatting on staged filed and include result in the commit
+
+Opt out through the `-n/--no-verify` flag available for most git commands

package.json

@@ -20,14 +20,20 @@
     "lint:ws": "pnpm dlx sherif@latest",
     "postinstall": "pnpm lint:ws",
     "typecheck": "turbo run typecheck",
-    "ui-add": "turbo run ui-add"
+    "ui-add": "turbo run ui-add",
+    "prepare": "husky"
   },
   "devDependencies": {
     "@acme/prettier-config": "workspace:*",
     "@turbo/gen": "^2.0.3",
+    "husky": "^9.0.11",
+    "lint-staged": "^15.2.7",
     "prettier": "^3.3.1",
     "turbo": "^2.0.3",
     "typescript": "^5.4.5"
   },
-  "prettier": "@acme/prettier-config"
+  "prettier": "@acme/prettier-config",
+  "lint-staged": {
+    "*": "prettier --ignore-unknown --write"
+  }
 }